Deliberate Practice: Why Reps Alone Do Not Make You Better

Picture someone who has driven a car nearly every day for twenty years. They are not a better driver than they were at year three. Somewhere back then they got comfortable enough, stopped paying close attention, and have been running on autopilot ever since. The hours kept piling up. The skill quietly stopped moving.

This is the uncomfortable thing about repetition. Doing something again is not the same as getting better at it. You can log a decade at the piano, or in a job, or on a tennis court, and arrive at the far end roughly as good as you were a few years in. The odometer says one thing. Your actual ability says another. What closes that gap has a name in the research, and it is not raw hours. It is deliberate practice.

The Violinists Who Practiced Alone

In the early 1990s, K. Anders Ericsson and two colleagues went to a music academy in West Berlin with a simple question: what separates the students headed for solo careers from the ones headed for the orchestra pit or the classroom? They sorted young violinists into three groups by ability, the best, the good, and those training to become music teachers, then had each one reconstruct how many hours they had spent practicing alone, year by year, since they first picked up the instrument.

The totals lined up almost too neatly. By age eighteen, the best violinists had practiced alone for an average of 7,410 hours. The good violinists: 5,301. The future music teachers: 3,420. More skill, more solitary practice, in that exact order, with no group out of place. Ericsson called it "complete correspondence" between skill level and accumulated practice.

It is a seductive result, and worth being careful with. The study could not prove the hours caused the skill. It leaned on people's memories of practice from a decade earlier, and memory flatters. But the shape of it was striking enough to launch a thousand arguments and one famous book.

Deliberate Practice Is Not Supposed to Feel Good

Here is where the word deliberate earns its keep. Ericsson was not describing hours logged. He was describing a specific, uncomfortable activity: work aimed squarely at what you cannot yet do, broken into pieces, repeated with full attention, and corrected by feedback the moment it goes wrong. The violinist does not play through the piece she already owns. She isolates the four bars that keep tripping her, slows them to a crawl, and drills them until the mistake stops, then hunts for the next weak spot.

That last part matters more than it sounds. Repetition is also how habits form, and once a skill turns automatic it runs without you, which we covered in how to build habits. Automaticity is a gift for the things you want on autopilot. It is the exact mechanism that stalls you on the things you are still trying to master. The reps that improve you are the ones that expose what you are bad at, which is precisely why most people quietly route around them.

The Rule Ericsson Never Wrote

You have likely heard that it takes 10,000 hours to master anything. That number comes from Malcolm Gladwell's 2008 book Outliers, not from Ericsson. Gladwell took the violinist study, rounded up to the tidiest figure in it, and turned it into a rule: ten thousand hours and you are an expert. It made a great line. It also was not what the study found.

The 7,410-hour figure was for age eighteen, not a finishing line. Ericsson's ten-thousand number was an extrapolation for the single best group carried out to age twenty, an average with enormous spread around it, not a threshold anyone had to cross. He spent years correcting the record. In his 2016 book Peak, written with Robert Pool, he put it as plainly as a scientist can.

The trouble with a round number is that it hides the two questions that matter: what you do inside those hours, and how much of the gap between people the hours explain at all. On the second question, the news is humbling.

Most of the Gap Is Unexplained

In 2014, Brooke Macnamara and two colleagues pooled every study they could find that measured practice and performance together, across music, games, sports, education, and the professions, then asked how much of the difference between performers the amount of practice actually accounted for.

About twelve to fourteen percent. That leaves the bulk of the difference between a great performer and an ordinary one sitting outside the practice column. It is tempting to fill that space with the word talent, but that would be a mistake. The unexplained part is a grab bag: when you started, who taught you, how motivated you were, how performance was measured, plain luck. Practice is a real contributor. It is simply not the whole story, or even most of it.

And how much it contributes swings wildly depending on where you are practicing. In predictable, rule-bound worlds, practice buys you a great deal. In messy ones, it buys you almost nothing.

The pattern is not subtle. Chess, scales, free throws: the rules hold still, so grinding at your weak spots pays off in a way you can count on. School and work are moving targets, where the thing you are judged on keeps shifting and a hundred other factors crowd in. There was even a tell in the method: studies that measured practice with loose retrospective interviews reported bigger effects than studies using actual practice logs. The softer the measurement, the larger practice appeared, a quiet warning about how much the original violinist study rested on memory. The authors' own conclusion was careful and worth repeating: deliberate practice is "important, but not as important as has been argued."

When They Ran the Violinist Study Again

Macnamara did not stop at the meta-analysis. In 2019, she and Megha Maitra rebuilt Ericsson's original violinist study from the ground up, this time pre-registered and double-blind, the safeguards that keep a researcher's hopes from leaking into the result. Same academy design, same three skill groups, same reconstruction of practice hours.

The clean staircase did not reappear. The best and the good violinists did not differ significantly in how much they had practiced. Practice still mattered, and the overall effect was substantial, but it was considerably smaller than in the 1993 original, and it no longer sorted people tidily by skill. None of this makes practice a myth or Ericsson wrong to have studied it. It means the real relationship is weaker and messier than the famous version, which is what tends to happen when a striking early finding meets a stricter method.

How to Practice So the Hours Count

None of this is a reason to practice less, but to practice with more intent, because the evidence keeps pointing at the same thing: raw repetition is the weakest version of the activity. If you want the hours to change you, a few moves separate practice from mere reps.

Work at the edge of what you can do, not the middle. The comfortable rep, the drill you could run in your sleep, is where improvement goes to idle. Aim instead at the specific thing you keep getting wrong, and stay there until it stops being the thing you get wrong.

The same move works for skills with no scales or drills. A writer who wants stronger openings does not write another whole essay. She writes ten first paragraphs for the same piece and compares them. A programmer who keeps making one kind of bug rebuilds that one function until the bug stops appearing. A language learner who fumbles one verb tense drills that tense, not the words she already knows. Writing and coding sit closer to school and work, where the meta-analysis found practice explained the least, so a narrow target matters even more there.

Then get feedback fast enough to correct the next attempt, not next month. Practice without feedback is just careful rehearsal of your current mistakes, and feedback is a skill of its own, easy to use badly, which we dug into in how to use feedback. The loop is short: try, see what broke, adjust.

You do not need a teacher to close that loop, though a good one makes it faster. What you need is some way to see the mistake soon after you make it. Record the four bars and listen back. Run the code and read the error. Say the sentence aloud next to a native recording and hear where yours drifts. A coach is one source of feedback, not the only one. The bigger risk of working alone is quieter: with nobody pointing at your weak spot, you slide back to the parts you already play well.

And hold the hours humbly. Ten thousand of them guarantee nothing, and in some domains they explain almost none of the difference between people. That is not a reason to quit. It is a reason to spend your practice where it actually changes something, on the specific weaknesses you can name, rather than banking time and waiting for mastery to accrue. It is also the version worth tracking. Future You lets you break a skill into the specific weak points you are working on and log the feedback on each, so your goal tracking shows where you actually got better, not how long you spent. It is free on iOS and Android.

What the Odometer Cannot See

The story we like about mastery is the montage: hours stacking up, the counter climbing, competence arriving on schedule. The research tells a plainer one. Most repetition changes nothing. A narrow slice of it, the effortful, feedback-corrected slice aimed at what you cannot yet do, is what quietly moves you forward, and even that operates inside limits worth respecting.

So the question to bring to your next practice session is not how long. It is whether you are working on something that is hard for you right now, and whether anything is telling you if you are getting it right. Get those two in place and the hours start to count. Skip them, and you can practice for twenty years and stay where you began, still no better than year three.